Drift
A substance-level difference between a delivered teaching and its source material, such as a dropped qualifier, a misstated timing, or a skipped step, distinguished from acceptable delivery variation, which changes wording or examples without changing what a listener actually learns to do.
Defined in 7 GAGE programs, which carry 15 distinct definitions of it. The wording above is taught in Senior AI Scam Defense.
How each discipline defines it
The same term does different work depending on who is using it. These are the definitions as each program teaches them, unedited.
A substance-level difference between a delivered teaching and its source material, such as a dropped qualifier, a misstated timing, or a skipped step, distinguished from acceptable delivery variation, which changes wording or examples without changing what a listener actually learns to do.
The gradual degradation of an AI model as the real world moves away from the data the model was trained on, so a model that was accurate at launch becomes quietly wrong over time. It is the technical reason an AI capability must be tended continuously rather than shipped and abandoned.
A bias that changes over time, driven by temperature variation, supply voltage changes, or component aging. Drift is distinguished from static bias by its dependence on an external variable. Gyroscope temperature drift is a primary limitation of pure inertial dead reckoning.
The gradual degradation of an AI capability as the real-world conditions it operates in diverge from the conditions it was built or trained on. Drift makes the Maintain bucket non-optional; a maintenance estimate near zero assumes the world stops changing. (see Topic 4.6)
Degradation as the world moves away from the data a system was fitted to. Detected through input distribution, which needs no model access, or through the input to outcome relationship, which needs outcome data and usually a delay.
Where it is taught
The exact lessons this term appears in. The first module of every program is free with a free account.
- Monitoring and Reviewing AI Agent Outputs · Agentic AI and Workforce Integration, Certified AI Practitioner: Workplace Foundations
- The taxonomy of data incidents: breach, leak, poison, drift, and wrongful training · Data Incidents, Certified AI Data Governance Professional (CADGP)
- Fixing the model, breaking it again: why fixes are never free · Build Before You Govern, Certified AI Governance Professional (CAIGP)
- Build, buy, or wrap: the decision framework with your organization's real budget · Shipping AI and Surviving the Incident, Certified AI Governance Professional (CAIGP)
- Analytics, Telemetry and Observability · Technology, Platforms and Vendors, Certified AI Transformation Professional (CATP)
- AI Risk Management · Governance and Responsible AI, Certified AI Transformation Professional (CATP)
- Building an AI-First Culture · Change Leadership and Politics, Certified AI Transformation Professional (CATP)
- AI systems and the enterprise landscape · AI Governance, AI Risk, AI Security, and AI Assurance, Certified Enterprise Governance, Risk and Compliance Professional (CEGP)
- The AI risk taxonomy · AI Governance, AI Risk, AI Security, and AI Assurance, Certified Enterprise Governance, Risk and Compliance Professional (CEGP)
- AI lifecycle controls · AI Governance, AI Risk, AI Security, and AI Assurance, Certified Enterprise Governance, Risk and Compliance Professional (CEGP)
Where it shows up at work
A definition is where Drift starts. The lessons above teach it inside the competencies below, which are what job descriptions ask for and what the GAGE credential records. Every link here is drawn from the same careers map that builds a role's readiness check.
Designs decision rights, committees, intake, approval tiers and escalation so routine uses move and consequential uses get reviewed.
2 lessons on this page teach it as a primary outcome.
Tiers vendors by use and impact, requests evidence instead of promises, tests in the customer's context, and plans monitoring and exit.
2 lessons on this page teach it as a primary outcome.
Turns controls into enforceable obligations: data use, change notice, audit rights, incident duties, subcontractors, IP, exit and deletion.
2 lessons on this page teach it as a primary outcome.
Roles that hire for it
Each role page shows the full competency set, what the seat decides, and a readiness check against your own evidence.
- Chief Compliance Officer, AI compliance editionneeds it at depth
- AI Governance Managerneeds it at depth
- Chief AI Officerneeds it at depth
- Chief Information Security Officer, AI security focusneeds working knowledge
- Chief Risk Officer, AI risk editionneeds working knowledge
- VP of AI Governance and Ethicsneeds it at depth
How job descriptions say it
Postings rarely use the textbook word. These are the phrases the careers map reads as asking for the competencies above.
Where the frameworks place it: NIST AI RMF GOVERN; ISO/IEC 42001 Clause 5; NIST AI RMF GOVERN 6 and MANAGE 3; ISO/IEC 42001 Annex A supplier controls.
One interview question about it
Sketch the governance operating model you would set up for a company deploying its first customer-facing AI. Who decides, who reviews, and who can stop it?
Asked of candidates for ai governance operating model design. Answering it in your own words is exactly how a GAGE topic is passed, and the credential records the pass where an employer can check it.
If you come from another field
Some backgrounds already carry this competency, and the readiness check gives credit for it.
- Privacy and data protection carries it as a claim to back up: Processor and vendor review is routine privacy work.
- Cybersecurity and IT carries it as a claim to back up: Supplier security assessment is routine.
- Nonprofit program and grants oversight carries it as a claim to back up: Reviewing grant and vendor agreements for data terms is vendor review.
Terms it appears with
Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.